simpleblog3 0.9.8

A simple Python 3 blogging system.

Simpleblog is a simple Python 3 blogging system. I use it to write and publish my own blog at http://blog.peterdonis.com. I wrote it because I couldn’t find an existing blogging system that made it sufficiently easy to write, format, and publish my blog.

My chief goal with simpleblog is for the system to stay out of my way; I want to be able to add features easily, but other than when I’m actually doing that, I want simpleblog to “just work”, so I don’t even have to think about it at all. That way I can think about what I’m writing instead. With the existing systems I’ve tried, I have ended up spending too much time figuring out the internals of the system in order to get things the way I want them. Admittedly, I have not tried many existing systems; but what I have read about the ones I haven’t tried has not encouraged me that any of them would work any better for me. So here we are.

If you just want to start using simpleblog, without digging into its internal details, then once you’ve installed it, you can copy the contents of one of the example blogs to a directory of your choice, and start writing your blog there. The layout of the example blogs, and the files in them, will give you a start. Before writing any entries, you will want to at least edit the blog.json or blog.yaml file to customize your blog’s metadata, and the template files in the templates subdirectory, which give extremely plain HTML pages by default.

Note that in order to use simpleblog, you will need to have installed plib3.stdlib (my library of useful Python 3 stuff, which is used in a number of places in simpleblog). It is available at http://pypi.org/project/plib3.stdlib. If you want to use YAML instead of JSON for your config and blog metadata files (I certainly find YAML much easier to type since I hate typing delimiters, as you will know if you read my blog), you will also need to have installed PyYAML, the YAML parsing library for Python (which in my opinion should be in the Python standard library).

Note: simpleblog3 is the Python 3 version of simpleblog. If you are using Python 2, see https://pypi.org/project/simpleblog.

Simpleblog’s Architecture

The structure of simpleblog is simple (no, that wasn’t intended to be humor, it’s just the way it naturally came out). There are five core object types: the config, the blog, pages, containers, and entries. The config lets you define or customize the internal behavior of the code, and all the other objects have a reference to it. The other object types fall into a simple hierarchy:

• The blog contains one or more pages, plus metadata which can be specified in a separate file from the config file; the default filenames are blog.yaml (or blog.json) and config.yaml (or config.json), but other filenames can be passed on the command line to the simpleblog-run script–see below;

• Each page wraps a “source”, which can be either a single entry, or a container;

• A container contains one or more entries that have something in common;

• An entry is the actual content.

It’s important to note that the above is all that the core objects implement, and it is completely general. Everything specific, such as what actual “sources” there are, which entries are in which containers, etc., is all defined in extensions. (Strictly speaking, there is one default container in every blog, which simply contains all its entries, and every blog has an index page, which uses that container as its “source”, plus a page for every individual entry. But that’s all that is in the blog by default. Of course, that by itself is enough to have a simple blog, which is part of the point.)

Templates

Simpleblog uses Python’s built-in string templating and formatting to render entries and pages. The example blogs illustrate the basics of how this works. This is one area where I do not have any items on my To Do list; the various fancy templating engines out there have their uses for highly dynamic web applications, but for a simple blog they are, in my opinion, extreme overkill. But the extension mechanism is there for anyone who disagrees and wants to use their favorite templating engine.

Extensions

Extensions allow pretty much every behavior of the four blog object types–everything above except the config–to be changed, and even allow new behaviors to be added. (I say “pretty much” only because I can’t be absolutely positive I have allowed for every possibility; but that’s my goal.) This is done with a simple (yes, you’ll see that word cropping up a lot…) but powerful mechanism. You write a Python module that contains a subclass of the BlogExtension class, and implements your desired changed or added behaviors, and add its name to the list of extensions in your config file. That’s it. Or, of course, you can use one of the extensions that come with simpleblog, listed below. I use all of them for my blog. They give good examples of how the extension mechanism can be used.

(Note: Strictly speaking, since extension names will be looked up as Python module names, they must be valid identifiers, which means they can’t include hyphens. However, simpleblog allows you to use hyphens when referring to extensions, as in the render-markdown extension; it converts the hyphens to underscores before looking up the module name. Command names are handled the same way–see below.)

• The archives extension adds containers for entries that were published during specific time periods–years, months, and/or days, depending on the config settings–and adds archive pages to the blog for each container.

• The categories extension allows you to classify entries by category, and adds a container and an index page for each category.

• The copyright extension automatically generates copyright metadata based on the starting and ending year of blog entries.

• The feed extension generates feeds for your blog’s index page. Both RSS 2.0 and Atom feeds are supported. This extension also supports archived feeds per RFC 5005 (this only works for Atom feeds since the RSS spec does not appear to support this), which lets you limit the size of your syndication feed file by archiving old entries.

• The folding extension allows your entries to have “short” versions that can appear in index pages, with links to the entry page that shows the entire entry (including the part “below the fold”).

• The grouping extension allows entries on index pages to be grouped, so that group headers and footers can appear in addition to the entries themselves. The default is to group by date, which goes along with the default sorting of entries in all containers, which is reverse chronological; but these can be changed by config settings (of course they should both be changed consistently).

• The indexes extension adds index pages to your blog that give links to all entries in either alphabetical (by title), chronological, or “key” (meaning the unique key assigned to each entry) order.

• The links extension allows you to add links to the previous and next entries in your blog’s containers to each entry. By default it only does this on single-entry pages, but this can be configured; also, which links actually appear on the page is controlled by a template you provide.

• The localize extension is currently experimental; all it does is add a “locale” config setting if certain other config settings are present. More localization functionality is on the To Do list; currently simpleblog is only tested with English ASCII text.

• The paginate extension allows splitting sources with many entries into multiple pages.

• The quote extension adds quoted versions of all URLS found in the blog’s metadata. I added this because I link to the W3C HTML validator for my blog’s index page, which wants quoted URLs, and this was an easy way to avoid having to type them into my blog metadata by hand. :)

• The render-markdown extension allows your entry source to be plain text using Markdown syntax; the extension then renders it into HTML. (Without any extension changing the rendering, simpleblog just uses your entry source unchanged as its rendered HTML.) There are config options to specify the output format for Markdown (the default is HTML 4) and to “pretty print” the output.

• The tags extension allows you to add tags to your entries, and adds a container and index page for each tag. This extension uses the caching mechanism for entry metadata (see below).

• The timestamps extension uses the caching mechanism to store immutable file timestamps. (Without any extension, an entry’s timestamp is the last modified time of its source file, but this means if you make any change at all to an entry once it is published, its time stamp changes, which may change where it appears in index pages.)

• The timezone extension makes entry timestamps timezone-aware (without this extension they are “naive” datetime objects). The timezone_name config setting lets you explicitly declare your blog’s timezone; otherwise your system’s local time zone setting will be used (note, however, that the utc_timestamps config setting can force the timezone to UTC; see notes in the change log). This extension requires the pytz library.

• The title extension allows you to specify a title for each entry in the entry’s source file. (Without any extension, the title of an entry is the same as its relative file name or URL path, which is probably not what you want.) It also supports very simple italics and bold formatting in the title.

Note that in some cases the order in which extensions are declared in your config file matters. The order in which extensions are listed in the config determines the order in which they are loaded, which determines the order in which they get to process whatever data they are processing, which can obviously make a difference if multiple extensions process the same data. The cases you are most likely to encounter are extensions that process the raw entry source data (the title, tags, and folding extensions all do, and the ordering that is known to work is the order in which I just gave them), and extensions that add sources in the form of new containers (the archives, categories, and tags extensions) vs. extensions that need to know all the containers in the blog (the links extension is the key one, and needs to be loaded after the ones listed just now).

Entry Metadata Caching

Entry metadata is often useful for putting entries into containers and ordering them properly. It is nice to be able to do this without having to actually ask the filesystem for any data on individual entries, by either statting or opening and reading the entry source files. Simpleblog provides a caching mechanism for entry metadata to make this simple. Just use the cached decorator on any property that represents metadata you want cached, and provide the name of the file the cache should be stored in.

Commands

All of the above is nice, but in order to actually use it, you have to have some kind of front end. The simpleblog-run script provides one. If run without any command at all, the script simply puts you into the Python interactive shell, with the simpleblog package loaded; I find this extremely useful for testing and debugging. But the script can also be enhanced with commands, by a mechanism similar to the extension mechanism.

(Note: As with extension names, hyphens in command names are converted to underscores before looking up the module, so you can use hyphens, as is done below, if you find them easier to type, as I do.)

• The publish command publishes your statically rendered blog via SSH to a remote host that will serve it. By default it uses the rsync command, but a config setting allows you to change the command name (though it must be a command that uses the same command-line syntax as rsync, such as scp). You can also configure the command options and the SSH user, the remote hostname, and the path on the remote host to publish to.

• The render-static command renders static versions of all the pages in your blog. A config setting controls the directory that the files are rendered to. For my blog, this is currently sufficient, since I publish it as static files.

• The serve-local command serves your statically rendered blog on localhost for testing. You can use command-line options to change the host name (or IP address) and port used (the defaults are localhost on port 8000), for example to allow testing on a LAN. Since the built-in Python SimpleHTTPServer is used, it is not recommended to try to serve your blog to the Internet using this command.

For quick help on usage, use the --help option to the simpleblog-run script. If a command name is provided, help specific to that command will be shown; otherwise, general help will be shown.

User-Defined Commands and Extensions

Simpleblog supports defining your own commands or extensions, separate from the ones supplied with simpleblog itself. All you have to do is set the command_dir or extension_dir config and supply Python modules that match the command or extension name you want to use. The command and extension loading mechanism will look in your user-defined directories first, so you can even define a command or extension with the same name as a pre-packaged one, and it will take precedence.

To Do

Add fancier example blogs to show how the various extensions work.

Add documentation other than this README file, both for users and for developers.

Add support for comments while still allowing the blog to be statically generated.